476 CHEMICAL ENGINEERING
clay substance, Al203-2Si02-2H2O. The clay therefore consists essentially of 258 (corresponding to molecular weight of the clay substance) parts, by weight, of clay substance and 4.97 X 60 = 298.2 of free silica or sand.
The mineralogical structure of the heat-resisting materials is of great importance hi the study of this subject and the application of the petrographic microscope both to the examination of thin sections and of fine grains is gaining in recognition more and more. It is impossible without this aid to aquire a thorough knowledge of the conversion of quartz to cristobalite and tridymite in silica brick, the formation of silliman-ite in clay firebrick, the crystallization of spinels, the recrystallization of periclase in magnesite brick, the formation of accessory minerals, etc. Both the chemical and mineralogical composition of refractories will be considered in dealing with the several types of refractories.
Refractoriness.—It is evident that a refractory 'retains its usefulness only as long as it continues to be sufficiently rigid to support its own weight and such loads as may be imposed upon it in furnaces.
Unfortunately, the term "refractoriness" has no specific meaning. Generally, it is supposed to represent the so-called melting point of a material. This confronts us with a serious difficulty due to the fact that the substances used for work of this kind possess no melting point in its proper physical conception. It is a well-known fact that silicates, like clays, lacking in well-defined crystalline structure and of high molecular viscosity, offer no definite transition point from the solid to the liquid phase, nor any other criterion, corresponding to the transformation of a substance from the anisotropic to the isotropic state. We are compelled therefore to depend upon deformation data, such as the rounding of edges, the bending of specimens of standard size and shape in the manner of Seger cones, or the formation of drops. But even when a standard condition of deformation has been established it is evident from theoretical considerations that the rate of heating will have a decided influence upon the deformation temperature, rapid firing resulting in a higher fusion point than that with a slow rate of heating. This condition is made still more complicated by the heterogeneous nature of the materials to be tested, which consist of particles of different substances varying greatly in size. It is thus seen that the determination of the softening temperature is not so simple as it might appear. Fortunately, the higher the temperature involved, the less marked seems to be the influence of the rate of heating upon the deformation point.
In making a softening point determination of a refractory it seems desirable to grind the specimen so that the material will pass the 80-mesh sieve and to mold from it, by the use of dextrin or other organic glue, small tetrahedra, % in. high and Y± in. wide at the base.
It is hardly necessary to point out that any metallic iron introduced by the grinding should be removed by means of a magnet. The tetrahedral specimens are placed on a plaque, made from a mixture of kaolin and sintered or fused alumina in the proportion of 1:1 in the case of clay and silica refractories. For basic specimens a mixture of sintered and low-fired magnesite, or of pulverized chromite, may be employed. Occasionally crushed carbon is used cemented together with tar or other organic cementing substances. The test cones are placed in position, alternating with the standard pyrometric cones (manufactured by Prof. Edward Orton, Columbus, Ohio). Usually three or four numbers are employed, representing as many softening points.
The object of the test then is to raise the temperature of the furnace to a point at which the specimen deforms by bending until its apex touches the plaque or in the case of certain refractories shows distinct evidence of fusion. The softening point of one of the standard cones indicated by the prescribed degree of bending establishes